EP3516466B1 - Procédé et système de détection d'attaques sur des systèmes physiques surveillés - Google Patents

Procédé et système de détection d'attaques sur des systèmes physiques surveillés Download PDF

Info

Publication number
EP3516466B1
EP3516466B1 EP17783590.7A EP17783590A EP3516466B1 EP 3516466 B1 EP3516466 B1 EP 3516466B1 EP 17783590 A EP17783590 A EP 17783590A EP 3516466 B1 EP3516466 B1 EP 3516466B1
Authority
EP
European Patent Office
Prior art keywords
signal
signals
noise
fake
monitored physical
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
EP17783590.7A
Other languages
German (de)
English (en)
Other versions
EP3516466A1 (fr
Inventor
Yevgeni NOGIN
Itay Baruchi
Charles Tresser
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Aperio Systems 2020 Ltd
Original Assignee
Aperio Systems 2020 Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Aperio Systems 2020 Ltd filed Critical Aperio Systems 2020 Ltd
Publication of EP3516466A1 publication Critical patent/EP3516466A1/fr
Application granted granted Critical
Publication of EP3516466B1 publication Critical patent/EP3516466B1/fr
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B23/00—Testing or monitoring of control systems or parts thereof
    • G05B23/02—Electric testing or monitoring
    • G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
    • G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
    • G05B23/0221—Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
    • H04L63/1408—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
    • H04L63/1408—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic by monitoring network traffic
    • H04L63/1416—Event detection, e.g. attack signature detection
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L63/00—Network architectures or network communication protocols for network security
    • H04L63/14—Network architectures or network communication protocols for network security for detecting or protecting against malicious traffic
    • H04L63/1433—Vulnerability analysis

Definitions

  • the present invention is about detection of forgery in data coming from monitored physical systems, such as: industrial control systems in power plants and manufacturing plants, the flight instruments of an aircraft or a ship, the flight control system of a remote-controlled missile, medical monitoring devices used during surgery in hospitals.
  • monitored physical systems such as: industrial control systems in power plants and manufacturing plants, the flight instruments of an aircraft or a ship, the flight control system of a remote-controlled missile, medical monitoring devices used during surgery in hospitals.
  • the data coming from those systems has critical importance for real time emergency decision making, and thus the assurance of its authenticity is of paramount importance.
  • industrial control systems such as the systems installed in many manufacturing plants, power plants and electricity distribution systems are sensitive to malicious attacks. These systems include means for detecting operational problems and malfunctions in the various components controlled by the control system. However, these systems are not immune against attacks on the control systems themselves.
  • the state of a physical system is controlled using various sensors for sensing the condition of the various components included in the system. Such sensors include direct measurement of voltage, current, power etc. and indirect measurements such as temperature vibration, sound etc.
  • Fig. 1 is a high level block diagram of an industrial control system as known in the art.
  • An industrial control system 100 may include a 3 levels structure.
  • a first device level that may include the various components and physical devices such as pumps 153, variable-frequency drives (VFDs) 154, sensors 152, valves 151, and other components such as motors, turbines, compressors and the like.
  • a second control level may include programmable logical controllers (PLC) 131 and/or remote terminal units (RTU) 132 for controlling the physical devices of the first device level.
  • a third supervisory level may supervise the controllers of the control level and may include a supervisory control and data acquisition ( SCADA ) 110.
  • SCADA 110 may have several interfaces that most often comprise human-machine interfaces (HMI) either as standalone HMI's 120 and/or through a control center at 121.
  • the data collected by the SCADA system is typically stored in a database known as the historian database.
  • Malicious attacks on a monitored physical system may include: stopping production processes, generate damage to the system and the like. Attackers may penetrate such systems and control them in a variety of ways, for example, taking control of the SCADA, changing the set points of the system (or a component of the system) in a way which will damage or stop their activity while optionally, overriding or canceling the alarms and protection logic in the SCADA.
  • Injection attack whereby the attacker changes the set point of the system while transmitting synthetic signals of standard operation state to the HMI
  • Transform attack whereby the attacker applies a transform such as multiplication by a constant on the set points of the system and a reverse transform on the monitoring control signals.
  • the attack can take form as any type of function including simple functions such as multiplication and division, and complex functions that will generate a smooth transition.
  • WO03091911 describes a system for detecting change in a data stream which comprises a distribution maintenance engine, a difference determining means and an alert generation engine.
  • the system detects change in the alert stream by the distribution maintenance engine maintaining a short term distribution that models the data stream and maintaining a long term distribution that models the data stream.
  • the difference determining means determines the difference between the short term distribution and the long term distribution.
  • the alert generation engine applies a statistical measure to the difference and generates an alert if the measure of the difference exceeds a threshold.
  • Some aspects of the invention may be directed to a computer system and computer method of detecting attacks on physical systems.
  • the system may include one or more databases and one or more controller configured to execute instructions.
  • the instructions may include the following method steps: receiving at least one signal related to a monitored physical system; extracting from the signal, a smooth portion of the signal; detecting one or more states of the monitored physical system by analyzing the smooth portion of the signal; obtaining a noise portion of the signal by subtracting the smooth portion from the signal; classifying the noise portion to produce one or more classified noise portions; determining one or more expected states of the system based on respective one or more classified noise portions; comparing the one or more expected states to the detected one or more states; and detecting an attack on the monitored physical system based on the comparison.
  • classifying the noise portion may include comparing the noise portion to classified fingerprints of noise portions stored in a database. In some embodiments, the method may further include dividing the noise portion into discrete segments and therefore, classifying the noise portion may include classifying at least some of the discrete segments. In some embodiments, classifying the noise portion may include obtaining noise signal features from the desecrate segments.
  • the at least one signal related to a monitored physical system comprises at least one of: temperature, current, voltage, pressure, vibrations, strains, power, phase, and loads.
  • the method may further include receiving a second signal, the second signal associated with a known state of the system; extracting a smooth portion of the second signal and obtaining the noise portion of the second signal; associating each one of the smooth portion of the second signal and the noise portion of the second signal with the known state of the system and saving the noise portion of the second signal in association with the known state of the system in the database.
  • the method may further include saving the smooth portion of the second signal in association with the known state of the system in a database and/or generating a classification module based on the associated smooth and noise portions.
  • the method may include updating the classification module with new associated smooth and noise portions.
  • the method may include receiving a set of signals related to a monitored physical system; identifying some of the signals in the set as non-fake signals; identifying at least one of the signals in the set to be fake signal; generating a first estimated signal for each of the at least one of the fake signal based on the identified non-fake signals; comparing the first estimated signal and the at least one fake signal; and detecting an attack on the monitored physical system based on the comparison.
  • the set of signals may include two or more signals related to two or more different parameters of the monitored physical system.
  • generating an estimated signal may include generating a specific vector based on a regression model for each of the at least one fake signal based on the identified non-fake signals.
  • the method may further include identifying at least one signal from the set of signals as a potentially fake signal based on the compression between the first estimated signal and the at least one fake signal; receiving one or more external signals from one or more sensors external to the monitored system; generating a second estimated signal for each of the at least one potentially fake signal based on the one or more external signals; comparing the second estimated signal and the at least one potentially fake signal; and detecting an attack on the monitored physical system based on the comparison.
  • the terms “plurality” and “a plurality” as used herein may include, for example, “multiple” or “two or more”.
  • the terms “plurality” or “a plurality” may be used throughout the specification to describe two or more components, devices, elements, units, parameters, or the like.
  • the term set when used herein may include one or more items.
  • the method embodiments described herein are not constrained to a particular order or sequence. Additionally, some of the described method embodiments or elements thereof can occur or be performed simultaneously, at the same point in time, or concurrently.
  • Some embodiments of the invention may be directed to an independent system for detecting attacks on monitored physical systems.
  • Such an independent system may not be directly connected or controlled by the controller or the SCADA of the monitored physical system, therefore may be immune to external attacks.
  • Such a system may include an independent control system located externally and optionally remotely from the monitored physical system that may receive signals directly from the physical devices (components) of the monitored physical system and may detect based on the received signal, a deviation or change in a typical signal that may indicate that an attack was made on the system.
  • a method may include detecting fake signals or deviated signal by identifying a fingerprint associated with various signals received from measurement of parameters of the physical devices.
  • fingerprints which are due to electro-mechanical machine noise - for example the electromagnetic noise generated by an electric motor or generator, mechanical vibrations of a spinning turbine, detectors and sensor's measurement noise, transmission line noise, temperatures fluctuations, specific sounds a system generates - for example each pump can have a specific sound pitch, fluctuation in pressures measurements - for example, in cooling liquid circulation systems, and the like .
  • a fingerprint according to embodiments of the invention may include any measurable parametric behavior of the monitored physical system which is embedded in the signals it is generating either directly (for example current) or indirectly (for example sound, which may not be an intended measured parameter of the monitored physical system, but allows deducing other parametric quantities).
  • the fingerprint may be specific and unique for each system.
  • the fingerprint should also be correlated to the state of the system.
  • an electrical motor may have a unique sound pitch which changes with the motor rotation speed. The change of the pitch as function of motor rotating speed may also be unique and may also be used for fingerprinting.
  • a monitored physical system may have several signals measured in parallel - a fingerprint can be generated by the combination of unique features of each signal.
  • the fingerprints identification may be the basis for signal authentication methods according to embodiments of the invention. These methods may be based on a hierarchy of analysis methods that help one to reduce as much as possible the number of false positive alarms and to reduce the possibility of faking the overall system.
  • Methods according to embodiments of the invention may include building a fingerprints database for a specific monitored physical system.
  • the database may include fingerprints related to more than one monitored physical system.
  • the database may supply to the system historical (e.g., typical) fingerprints of the system to be compared to a real-time measurements.
  • a system 200 may include a processor 205 and one or more databases 250 and 440.
  • System 200 may include a machine learning module 245.
  • System 200 may be in communication with control system 100, for Fig. 1 , however may be completely independent from control system 100.
  • System 200 may receive signals and information from control system 100, but cannot be affected, changed, amended, decoded and/or receive instructions from control system 100.
  • System 200 may further receive information and signals from detection system 300 located and/or included in the physical devices of the monitored physical system.
  • Processor 205 may include a central processing unit (CPU), a chip or any suitable computing or computational device.
  • Processor 205 may include an operating system and a memory for storing thereon instructions and code for executing methods according to some embodiments of the invention.
  • Processor 205 may perform analysis of signal received from detection system 300, for example, by using authentication algorithm and performs required classification and regression analysis. The authentication algorithm and regression analysis may be performed in module 245 using fingerprints and regression models stored in databases 122 (history server) and 115. In some embodiments, module 245 may be included in processor 205.
  • System 200 may be remotely located with respect to the monitored physical system and may receive the signals from system 100 or system 300 by wired or wireless communication. Methods for building fingerprints databased are disclosed with respect to Figs. 4 and 6 .
  • a notification may be send to control system 100 and/or to a list of users that should to be informed.
  • system 200 may receive signals directly measured from monitored devices, additional sensors either internal or external to the monitored physical system that may be used by the authentication system.
  • Detection system 300 may increase the strength of the signal authentication and decrees the number of false alarms.
  • System 300 may include audio and vibration sensors, IR cameras, a variety of electromagnetic sensors and other types of short and long range detectors.
  • the signal from system 300 may be integrated in the authentication process as part of a multi-signal authentication algorithm.
  • Fig. 3 is a flowchart of a method of detecting attacks on monitored physical systems according to some embodiments of the present invention.
  • Embodiments of the method may be performed by a system such as system 200 or may any other suitable computation system.
  • Embodiments of the invention may include receiving at least one signal related to a monitored physical system, for example, from SCADA 110 or from history database 122, in operation 305.
  • a signal T i may be acquired by measurements of a physical state of the device (e.g., its temperature, current, voltage, pressure, vibrations, strains, loads, power, phases, and the like).
  • the signal may be sampled every ⁇ units of time, such that n ⁇ represents the time of sampling point #n.
  • t is used instead of n ⁇ .
  • Some of the main sources of noise in such a system are typically sampling noise added to the signal by the analog transmission line and analog to digital (AD) converter and activity noise of the device itself.
  • Embodiments of the invention may include separating (e.g., de-noising) the signal to extract a smooth portion of the signal, in operation 310.
  • the smooth (state) portion and a noise portion of the signal may be separated using one or more de-noising methods.
  • de-noising methods may include but are not limited to running average, exponential running average, filtering or any other methods known to those skilled in the art of signal processing or the art of statistics.
  • the smooth portion may be analyzed for detection of the different sates of the system, in operation 320.
  • Various methods can be used to descript system states, for example, using histogram, feature lists, unbiased estimation (Maximum likelihood, least squares, Bayesian estimators and deep neural networks), Density Estimation and Kernel estimation, V-optimal histograms be saved and the like.
  • a system state may include one or more set points of the system (e.g., predetermined working conditions, such as, power, temperature, pressure and the like), operation modes of the system (e.g., working or not-working during a time period) and/or measured values (e.g., of the smooth portion) of the monitored system parameter(s) (e.g., the temperature, pressure, vibration etc.) that are constant over a period of time (e.g., seconds, minutes, hours, days, etc.) or vary within predetermined margins (e.g., margins below a predetermined threshold value).
  • the smooth portion of the signal X i may be divided into discrete segments. The segments may be either overlapping or consecutive, and the specific state of the system may be identified.
  • the states of all segments for the i th signal may be collected into a vector si N where N is the number of segments.
  • the noise portion of the signal W i (t) may be obtained by subtracting the de-noised smooth portion -D i (t) from the original (total) signal T i (t), in operation 330.
  • D i (t) may be an approximation of X i (t), (having tradeoffs between quality and speed of computation).
  • W i (t) may be analyzed to obtain noise signal features.
  • the signal may be divided into discrete segments (same segments used to construct the de-noised signal D i (t)).
  • a variety of methods may be used for obtaining features, for example, Statically modeling of the noise using methods (e.g., statistical moments, L-moments and the like), Autocorrelation, ARMA, Spectral analysis of the noise using methods (e.g., calculating the energy value at different frequency bands using different types of transforms such as classical Fourier transform, Wavelets or wavelets packets transform and the like), Hilbert-Huang transform and Empirical mode decomposition (EMD), the Continuous time Short Time Fourier Transform (Continuous time STFT), the two-sided Laplace transform, the Melin transform, the S-Transform, the Wigner-Ville Distribution, Composite spectrum (CS), Stationary wavelet transform (SWT) a wavelet transform algorithm designed to overcome the lack of translation-invariance of the DWT and the like.
  • Other methods may include, State space features, fractal features and entropy feature such as - minimum embedding dimension, Lyapunov exponents, box dimension, fractal dimension, entropy and
  • the noise portion of the signal may be collected to a matrix FP i N - where N is the window number.
  • the noise features for the segment collected in to a vector may then be used for classifying the noise portion, in operation 340.
  • the noise portion (e.g., set in a vector) may be used as an input to corresponding classification model CM i , received from a fingerprint database (e.g. database 115), in operation 350.
  • CM i e.g. database 115
  • at least some desecrate segment of the noise portion may be separately classified.
  • the output of classification model may be an expected state of the system values (e.g., a smooth signal).
  • the expected state obtained from the calcification model may then be compared to the determined state of the system, at operation 360.
  • Fig. 4 is a flowchart of a method of building a fingerprints database for an monitored physical system according to embodiments of the present invention.
  • the method of Fig. 4 may be performed by any controlling or computing system for example, system 200.
  • Operations 405, 410, 420 and 430 may be substantially the same as operations 305, 310, 320 and 330 of the method of Fig. 3 .
  • signal X;(t) associated with a known state of the system may be received, in operation 405, from a specific monitored physical system (e.g., a power plant) when the monitored physical system works under "normal” or “expected” conditions, meaning that the system is not under any attack and all operation parameters and/or detected signals are as expected.
  • a specific monitored physical system e.g., a power plant
  • both features matrix FP i N (of the noise signal) and state vector s i N (fo the smooth signal) may be used as a training and validation set for a classification algorithm.
  • the received signal may be de-noised to extract a smooth portion of the signal and obtaining the noise portion of the signal, for example, according to operations 310 and 340 of Fig. 3 .
  • the smooth portion of the signal and the noise portion of the signal with the known state of the system may be associated with the known state of the system.
  • classification algorithms that may be used at this stage such as, support vector classification (SVCs), random forest and the like.
  • the classification algorithm (or algorithms) may be used to generate a classification model CM i .
  • the classification model for the signal indexed by i is then stored in a database (e.g., database 115), in operation 450.
  • the signal indexed stored in database 115 may be related to a specific monitored physical system working at ordinary or expected conditions.
  • Fig. 5 is a flowchart of a method of detecting attacks on monitored physical systems according to some embodiments of the present invention.
  • the method of Fig. 5 may be performed by system 200 or by any other system.
  • the method of Fig. 5 may be applied to signals defined as potential fake signals by the method of Fig.3 .
  • a set of signals T i (t) related to a monitored physical system may be received, in operation 505, for example, from SCADA 110 and/or history server 122.
  • at least two of signals T i (t) may be related to two different parameters of the monitored system.
  • T1(1) may be the temperature of the steam in the entrance to the turbine, T2(t) the pressure of the steam at the exit of the turbine, T3(t) the torque in the main axis of the turbine and T4(t) the produced electrical power.
  • the set of signals may be continuously monitored looking for fake signals by analyzing segments (the same size and overlap as used in the training stage). Time segments of all signals T 1 ...T n may be taken, in operation 510. Operation 510 may include at least some of operations 310-360 of Fig. 3 conducted on a plurality of signals.
  • At least some of the signals in the set may be authenticated in operation 515, using for example, the method disclosed in operation 365.
  • form the set of signals T 1 ...T n being analyzed some of the signals may be identified as non-fake or authenticated signals while at least one of the signals T j may be identified as being a fake signal or un-authenticated signal.
  • embodiments may include applying a corresponding regression model on the identified at least one un-authenticated signal T j in the set.
  • the authenticated (non-fake) may be used to further estimate the non-authenticated (e.g., fake) signal(s).
  • An estimated signal T ⁇ j may be generated, in operation 540, based on the authenticated signals (e.g., the 9 authenticated signals) from the same set.
  • two signals T j and T ⁇ j may be compared, in operation 560.
  • T ⁇ j the estimated value
  • T 1 the measured value (e.g., T1 can be either a fake signal or real signal) and if there is discrepancy alert.
  • the comparison may be done, for example, by mean of subtraction.
  • the absolute subtraction value is above a predetermined threshold the signal may unauthenticated. If the two match up to some predefined level, the system may continue its monitoring and training of the signal fingerprints. If the two do not match an indication of a fake signal may be generated and transmitted to the required users and an attack on the monitored physical system may be detected based on the comparison, in operation 565.
  • the method of Fig. 5 may further include further evaluation of at least one potentially fake signal identified based on the compression between the first estimated signal and the at least one un-authenticated signal in operation 560.
  • the method my further include receiving one or more external signals from one or more sensors external to the monitored system. For example, if the monitored system is a turbine in a factory, a microphone and/or IR camera may be placed outside a factory to measure sound and/or heat coming from the turbine.
  • the method may further include generating a second estimated signal for each of the at least one potentially fake signal based on the one or more external signals.
  • a regression model may be used to validate the some of the monitored system signals based on the external signals.
  • the method may include comparing the second estimated signal and the at least one potentially fake signal and detecting an attack on the monitored physical system based on the comparison.
  • Fig. 6 is a flowchart of a method of building a fingerprints database for a monitored physical system according to embodiments of the present invention.
  • the method of Fig. 6 may be performed by any controlling or computing system for example, system 200.
  • a set of signals T i (t) associated with a known state of the system may be received, in operation 605, for example, from SCADA 110 and/or history server 122.
  • the received set of signals may be also received from the monitored physical system when the system operates at "normal" operation, when the system is not under attack, and all the measured signals are expected signals.
  • the received signal may be collected when the system is under attack, and may be identified as "fake signals" for later use.
  • the set of signals may be continuously monitored by analyzing segments (the same size and overlap as used in the training stage). Time segments of all signals T 1 ...T n may be taken, in operation 610.
  • a clustering algorithm may be used to identifying a sub-set of correlated signals G i , in operation 620.
  • the correlation may be done in order to verify that signal received from the same physical system, for example, using either linear or non-linear methods.
  • a variety of clustering algorithms may be used at this stage, for example, hierarchical clustering, K-means, spectral clustering, Gaussian mixtures and the like.
  • Training of the regression model can also be performed every predefined time constant (such as every week or every month) or the training can be done continuously as each authenticated segment is then used to re-train the set of regression models.
  • the signals of each set G i may be used to build a regression model RM G , in operation 635.
  • the regression model may be stored in a multi signal model database in operation 645, to later be used in operations 530 and 540 of the method of Fig. 5 .
  • a variety of methods may be used to generate the regression model, for example, linear regression, Bayesian regression, support vector regression and the like. It should be noted that in building the regression model every possible combination of signals may be used so that each signal can be estimated even if only one of the other signals is authenticated.
  • training or updating the regression model may also be performed every predefined time constant (such as every week or every month) or the training can be done continuously as each authenticated segment is then used to re-train the set of regression models.
  • any form of model update (or combination of them) has been used, it will usually be needed to eliminate some of the former members of the training set: one may then for instance use a simple rule of replacement such as FIRST IN-FIRST OUT, LAST IN-FIRST OUT.
  • the methods and system disclosed above may be directed to recognize false signals and in particular attacks on an electricity production plant.
  • Embodiments of the method may include comparing noise segment of a signal received from one or more devices included in the electricity production plant to noise segments stored in a database.
  • the comparison may include at least one of: testing models for fake time signals, using one or more multi-class classification methods for identifying if a signal segment is more likely genuine or has more likely been built according to one or more methods to generate fake signals and using one or more one-class classifiers, built using one or more approach such as Support Vector Machines, density methods, extreme value theory.
  • the comparison may include, one or more one-class classifiers, built using one or more approach such as Support Vector Machines, density methods, extreme value theory.
  • the method may include receiving a plurality of signals, filtering the signals to receive smooth signals and comparing total smooth signals to each other.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Security & Cryptography (AREA)
  • Computer Hardware Design (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Testing And Monitoring For Control Systems (AREA)
  • Selective Calling Equipment (AREA)

Claims (14)

  1. Procédé informatique pour détecter des attaques sur des systèmes physiques, le procédé comprenant :
    la réception d'au moins un signal (305) lié à un système physique surveillé (200) ;
    le procédé informatique étant caractérisé par :
    l'extraction, à partir du signal, d'une portion lisse du signal ;
    la détection d'un ou plusieurs états du système physique surveillé en analysant la portion lisse du signal ;
    l'obtention d'une portion bruit du signal en soustrayant la portion lisse du signal ;
    le classement de la portion bruit pour produire une ou plusieurs portions bruit classifiées ;
    la détermination d'un ou plusieurs états attendus du système sur la base d'une ou de plusieurs portions bruit classifiées respectives ;
    la comparaison d'un ou de plusieurs états attendus à un ou plusieurs états détectés ; et
    la détection d'une attaque sur le système physique surveillé sur la base de la comparaison.
  2. Procédé informatique selon la revendication 1, dans lequel la classification de la portion bruit comprend :
    la comparaison de la portion bruit aux empreintes digitales des portions bruit classées stockées dans une base de données.
  3. Procédé informatique selon la revendication 1 ou 2, comprenant en outre :
    la division de la portion bruit en segments discrets,
    et dans lequel la classification de la portion bruit comprend la classification d'au moins certains des segments dégradés.
  4. Procédé informatique selon la revendication 3, dans lequel la classification de la portion bruit comprend en outre l'obtention de caractéristiques de signal de bruit à partir des segments discrets.
  5. Procédé selon l'une quelconque des revendications précédentes, dans lequel le au moins un signal lié à un système physique surveillé comprend au moins un parmi : température, courant, tension, pression, vibrations, déformations, puissance, phase et charges.
  6. Procédé selon l'une quelconque des revendications précédentes, comprenant en outre :
    la réception d'un deuxième signal, le deuxième signal étant associé à un état connu du système ;
    l'extraction, à partir du deuxième signal, d'une portion lisse du deuxième signal et l'obtention de la portion bruit du deuxième signal ;
    l'association de chacune de la portion lisse du deuxième signal et de la portion bruit du deuxième signal à l'état connu du système ; et
    la sauvegarde de la portion bruit du deuxième signal en association avec l'état connu du système dans la base de données.
  7. Procédé informatique selon la revendication 6, comprenant en outre :
    de sauvegarder de la portion lisse du second signal en association avec l'état connu du système dans une base de données.
  8. Procédé informatique selon la revendication 6, comprenant en outre :
    de générer un module de classification basé sur les portions lisses et de bruit associées.
  9. Procédé informatique selon la revendication 8, comprenant en outre ;
    la mise à jour du module de classification avec de nouvelles portions lisses et bruit associées.
  10. Procédé informatique selon l'une quelconque des revendications précédentes, comprenant en outre :
    la réception d'un ensemble de signaux liés à un système physique surveillé ;
    l'identification de certains des signaux de l'ensemble comme des signaux non faux ;
    l'identification d'au moins l'un des signaux de l'ensemble comme étant un signal faux ;
    la génération d'un premier signal estimé pour l'au moins un signal faux sur la base des signaux non-faux identifiés ;
    la comparaison du premier signal estimé et le au moins un signal faux ; et
    la détection d'une attaque sur le système physique surveillé sur la base de la comparaison.
  11. Procédé selon la revendication 10, dans lequel l'ensemble des signaux comprend deux signaux ou plus liés à deux ou plusieurs paramètres différents du système physique surveillé.
  12. Procédé selon la revendication 10 ou la revendication 11, dans lequel la génération d'un signal estimé comprend la génération d'un vecteur spécifique sur la base d'un modèle de régression pour chacun du au moins un signal faux sur la base des signaux non-faux identifiés.
  13. Procédé des revendications 10 à 12 comprend en outre :
    l'identification d'au moins un signal à partir de l'ensemble des signaux, comme un signal potentiellement faux sur la base de la compression entre le premier signal estimé et le au moins un signal faux ;
    la réception d'un ou plusieurs signaux externes d'un ou plusieurs capteurs externes au système surveillé ;
    la génération d'un deuxième signal estimé pour chacun du au moins un signal potentiellement faux sur la base d'un ou plusieurs signaux externes ;
    la comparaison du deuxième signal estimé et le au moins un signal potentiellement faux ; et
    la détection d'une attaque sur le système physique surveillé sur la base de la comparaison.
  14. Système (200) pour détecter des attaques sur des systèmes physiques, le système (200) comprenant :
    une ou plusieurs bases de données (250, 440) ; et
    un processeur (205) configuré pour exécuter les étapes du procédé selon l'une quelconque des revendications 1-13
EP17783590.7A 2016-09-21 2017-09-19 Procédé et système de détection d'attaques sur des systèmes physiques surveillés Active EP3516466B1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201662397362P 2016-09-21 2016-09-21
PCT/IL2017/051061 WO2018055616A1 (fr) 2016-09-21 2017-09-19 Procédé et système de détection d'attaques sur des systèmes physiques surveillés

Publications (2)

Publication Number Publication Date
EP3516466A1 EP3516466A1 (fr) 2019-07-31
EP3516466B1 true EP3516466B1 (fr) 2020-08-19

Family

ID=60081115

Family Applications (1)

Application Number Title Priority Date Filing Date
EP17783590.7A Active EP3516466B1 (fr) 2016-09-21 2017-09-19 Procédé et système de détection d'attaques sur des systèmes physiques surveillés

Country Status (4)

Country Link
US (1) US11190530B2 (fr)
EP (1) EP3516466B1 (fr)
IL (1) IL265499B (fr)
WO (1) WO2018055616A1 (fr)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11943236B2 (en) * 2018-04-26 2024-03-26 Hitachi Energy Ltd Technologies for detecting cyber-attacks against electrical distribution devices
US11306705B2 (en) * 2019-10-04 2022-04-19 City University Of Hong Kong System and method for monitoring a device
EP3879362A1 (fr) * 2020-03-11 2021-09-15 Siemens Gamesa Renewable Energy A/S Procédé mis en uvre par ordinateur pour identifier un accès non autorisé d'une infrastructure informatique de parc éolien
WO2021251906A1 (fr) * 2020-06-11 2021-12-16 Singapore University Of Technology And Design Procédé et système de détection d'anomalie dans un processus physique associé à un système de commande en réseau
US20220036238A1 (en) * 2020-07-30 2022-02-03 Tektronix, Inc. Mono channel burst classification using machine learning
CN115801411B (zh) * 2022-11-18 2023-05-16 国网河北省电力有限公司电力科学研究院 一种电力物联网络攻击行为的高阶数据特征提取和识别方法

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040059947A1 (en) * 2001-12-12 2004-03-25 Lee Susan C. Method for training a hierarchical neural-network intrusion detector
GB0209368D0 (en) 2002-04-24 2002-06-05 Neural Technologies Ltd Method and system for detecting change in data streams
US7313092B2 (en) * 2002-09-30 2007-12-25 Lucent Technologies Inc. Apparatus and method for an overload control procedure against denial of service attack
US7665135B1 (en) * 2005-06-03 2010-02-16 Sprint Communications Company L.P. Detecting and addressing network attacks
US7808916B1 (en) 2005-12-28 2010-10-05 At&T Intellectual Property Ii, L.P. Anomaly detection systems for a computer network
US9043894B1 (en) * 2014-11-06 2015-05-26 Palantir Technologies Inc. Malicious software detection in a computing system
JP6306739B2 (ja) * 2014-11-18 2018-04-04 日本電信電話株式会社 悪性通信パターン抽出装置、悪性通信パターン抽出方法、および、悪性通信パターン抽出プログラム
US9906544B1 (en) * 2014-12-02 2018-02-27 Akamai Technologies, Inc. Method and apparatus to detect non-human users on computer systems
US9571512B2 (en) * 2014-12-15 2017-02-14 Sophos Limited Threat detection using endpoint variance
US9749345B2 (en) * 2015-04-22 2017-08-29 International Business Machines Corporation Reporting security vulnerability warnings
US9948678B2 (en) * 2015-10-27 2018-04-17 Xypro Technology Corporation Method and system for gathering and contextualizing multiple events to identify potential security incidents
US10609053B2 (en) * 2015-11-24 2020-03-31 Intel Corporation Suspicious network traffic identification method and apparatus
US10805326B1 (en) * 2016-11-21 2020-10-13 Alert Logic, Inc. Systems and methods for threat visualization with signature composure, spatial scale and temporal expansion
US11095642B2 (en) * 2018-11-07 2021-08-17 Cylance Inc. Verifying user identity through human / computer interaction

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
None *

Also Published As

Publication number Publication date
US20200186549A1 (en) 2020-06-11
IL265499A (en) 2019-05-30
EP3516466A1 (fr) 2019-07-31
IL265499B (en) 2022-06-01
US11190530B2 (en) 2021-11-30
WO2018055616A1 (fr) 2018-03-29

Similar Documents

Publication Publication Date Title
EP3516466B1 (fr) Procédé et système de détection d'attaques sur des systèmes physiques surveillés
Wang et al. Prognostics and health management: A review of vibration based bearing and gear health indicators
US10594712B2 (en) Systems and methods for cyber-attack detection at sample speed
Elnour et al. A dual-isolation-forests-based attack detection framework for industrial control systems
US10819725B2 (en) Reliable cyber-threat detection in rapidly changing environments
US11170314B2 (en) Detection and protection against mode switching attacks in cyber-physical systems
Liu et al. Intelligent wind turbine blade icing detection using supervisory control and data acquisition data and ensemble deep learning
US10204226B2 (en) Feature and boundary tuning for threat detection in industrial asset control system
US10678912B2 (en) Dynamic normalization of monitoring node data for threat detection in industrial asset control system
CA2969414C (fr) Appareil et methode de surveillance d'un dispositif comportant une piece mobile
KR101955305B1 (ko) 희소 코딩 방법론을 활용한 가스 터빈 센서 고장 검출
Mosallam et al. Nonparametric time series modelling for industrial prognostics and health management
Filev et al. An industrial strength novelty detection framework for autonomous equipment monitoring and diagnostics
EP3847586A1 (fr) Procédé mis en oeuvre par ordinateur, produit-programme d'ordinateur et système de détection d'anomalie et/ou de maintenance prédictive
US11658988B2 (en) Dynamic physical watermarking for attack detection in cyber-physical systems
US11639966B2 (en) Enhanced electrical signature analysis for fault detection
US11880464B2 (en) Vulnerability-driven cyberattack protection system and method for industrial assets
CN113383160A (zh) 一种用于对风力涡轮机进行计算机实现的监控的方法
CN117076869B (zh) 旋转机械时频域融合故障诊断方法及系统
Purbowaskito et al. The potentiality of integrating model-based residuals and machine-learning classifiers: An induction motor fault diagnosis case
Elyassami et al. Bearing fault diagnosis and classification based on KDA and alpha-stable fusion
Benmoussa et al. Support vector machine classification of current data for fault diagnosis and similarity-based approach for failure prognosis in wind turbine systems
Mosallam et al. Unsupervised trend extraction for prognostics and condition assessment
Radhakrishnan et al. Complexity-entropy feature plane for gear fault detection
Liao et al. Machine anomaly detection and diagnosis incorporating operational data applied to feed axis health monitoring

Legal Events

Date Code Title Description
STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: UNKNOWN

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE

PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE

17P Request for examination filed

Effective date: 20190415

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

AX Request for extension of the european patent

Extension state: BA ME

RIN1 Information on inventor provided before grant (corrected)

Inventor name: BARUCHI, ITAY

Inventor name: NOGIN, YEVGENI

Inventor name: TRESSER, CHARLES

DAV Request for validation of the european patent (deleted)
DAX Request for extension of the european patent (deleted)
GRAP Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOSNIGR1

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: GRANT OF PATENT IS INTENDED

INTG Intention to grant announced

Effective date: 20200311

GRAS Grant fee paid

Free format text: ORIGINAL CODE: EPIDOSNIGR3

RAP1 Party data changed (applicant data changed or rights of an application transferred)

Owner name: APERIO SYSTEMS (2020) LTD.

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE PATENT HAS BEEN GRANTED

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

REG Reference to a national code

Ref country code: CH

Ref legal event code: EP

REG Reference to a national code

Ref country code: DE

Ref legal event code: R096

Ref document number: 602017022079

Country of ref document: DE

REG Reference to a national code

Ref country code: AT

Ref legal event code: REF

Ref document number: 1304671

Country of ref document: AT

Kind code of ref document: T

Effective date: 20200915

REG Reference to a national code

Ref country code: IE

Ref legal event code: FG4D

REG Reference to a national code

Ref country code: CH

Ref legal event code: NV

Representative=s name: VALIPAT S.A. C/O BOVARD SA NEUCHATEL, CH

REG Reference to a national code

Ref country code: LT

Ref legal event code: MG4D

REG Reference to a national code

Ref country code: NL

Ref legal event code: MP

Effective date: 20200819

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: NO

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20201119

Ref country code: GR

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20201120

Ref country code: BG

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20201119

Ref country code: LT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: HR

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: PT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20201221

Ref country code: FI

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: SE

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

REG Reference to a national code

Ref country code: AT

Ref legal event code: MK05

Ref document number: 1304671

Country of ref document: AT

Kind code of ref document: T

Effective date: 20200819

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: NL

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: RS

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: LV

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: PL

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: IS

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20201219

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: SM

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: EE

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: RO

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: DK

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: CZ

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

REG Reference to a national code

Ref country code: DE

Ref legal event code: R097

Ref document number: 602017022079

Country of ref document: DE

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: MC

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: AL

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: AT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: ES

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: SK

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

26N No opposition filed

Effective date: 20210520

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: IT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: SI

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: TR

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: MT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

Ref country code: CY

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: MK

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20200819

REG Reference to a national code

Ref country code: CH

Ref legal event code: PL

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: LI

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20220930

Ref country code: CH

Free format text: LAPSE BECAUSE OF NON-PAYMENT OF DUE FEES

Effective date: 20220930

PGRI Patent reinstated in contracting state [announced from national office to epo]

Ref country code: LI

Effective date: 20230623

Ref country code: CH

Effective date: 20230623

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: CH

Payment date: 20231001

Year of fee payment: 7

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: IE

Payment date: 20240702

Year of fee payment: 8

Ref country code: DE

Payment date: 20240702

Year of fee payment: 8

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: GB

Payment date: 20240701

Year of fee payment: 8

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: BE

Payment date: 20240703

Year of fee payment: 8

Ref country code: LU

Payment date: 20240913

Year of fee payment: 8

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: FR

Payment date: 20240702

Year of fee payment: 8